Setting this discussion up to gather up and brainstorm ideas.
LIST OF EQUIPMENT AND MATERIALS UNDER CONSIDERATION:
Materials to be Used:
Some design considerations:
Everyone feel free to inject any comments. This is your chance to provide input on what features go into the DIYDrones CNC Machine!
Current Bed Size:
Currently we're thinking the bed size will be 18x24 inches.
Insights from bGatti:
Some observations I can offer from experience:
1. Inexpensive skate bearing are not well sealed, and will likely stick if exposed to wood dust.
2. The most vibration on my machine is twist across the gantry - so make the gantry height ~twice the height of the lower extremity (from the table).
3. Parallel ports are increasingly harder to find - especially on laptops.
4. It doesn't take a lot of power to move a cnc machine - you will probably want to turn down the power to avoid breaking bits / bending the spindle shaft anyway - so don't specify a motor system which is stronger than your specified spindle. Anything which can be cut quickly on a strong machine, can be cut slowly on a lesser machine. $1000 is a lesser machine.
5. I have a wooden bed; I'd love an aluminum bed, but with $1000 to spend, the wooden bed isn't my biggest challenge. (It's gantry twist and probably runout for PCB Boards).
6. 3D printers want speed more than brute rigidity - most blogs on the subject of hybrids concede this point - which said - there are many parts which can be made slowly - but some parts call for a flying buttress - which does depend on a certain rate of speed.
Reading up on CNC machines lead me to read about 3D printing (a lot of things seem to lead me to read about 3D printing...)
It occured to me that maybe we could use a board from the RepRap family of open-source 3D printers to control our motors:
They're basically GCode interpreters after all. Right now I can only see them using steppers, but maybe we can tweak the firmware to use servos as well. Maybe someone with better electronics knowledge than me can investigate if the outputs can be fed to the UHU control boards. That way we could DIY the control board as well, if we took the Generation 7 board for example.
I also stumbled upon this seemingly simple hack for a 3D printer extruder, though I don't expect a CNC machine to do a very good job as a 3D printer... http://hackaday.com/2012/06/05/tearing-apart-a-hot-glue-gun-for-a-3...
We should keep the core project as CNC printer and people who are willing to can then try to gun for the extruder head.
Sounds great, Kevin. Even at single quantities, we seem close to the $60 per board that Monroe, was hoping for.
I've also added this new spreadsheet to the top of the discussion, so we can get too it easier.
The problems so far with 3D printing on a CNC are related to speed vs strength. Most CNC can't move fast enough.
Reading more about the RepRap electronics, they can't drive servos nor use anything else than USB (so no LinuxCNC). Better find something else.
Ok then, I'll bite ;)
What we'd need is a bowden extruder. Simple enough at the nozzle, it would allow us to keep most of the setup out of the way (maybe on the outer side of the gantry wall). It could be switched from another tool faster this way. Auto tool-changer compatible even?
You could use a regular electric frying pan and just mill the sides down with the machine and take the legs off :) ??? lol
And suddenly all the thrift stores wonder why the electric skillets are disappearing off their shelves...
There was a reference earlier to "parallel ports are not common". I would like to add that serial ports aren't so common anymore either. Fortunately, the CNC machines at our office each have an old serial port and I've been adapting each one to use the network for programming and even wifi. It's made the machines so much easier to use. Our 3D printer is also networked.
An embedded webpage for settings, etc... would be a nice plus.
Also, not that mobile access is needed for a CNC machine, but a networked / wifi CNC machine also opens up the possibility of some nice apps replacing the joystick and other type of controls.
Doesn't this work http://www.amazon.com/IOGEAR-GUC1284B-USB-Parallel-Adapter/dp/B0001...
I've used similar cable, when I wanted to connect a computer w/o parallel port to an old printer.
I, personally, am not against a printer port, but they are getting more and more difficult to find on computers. If we go with a printer port, I think we should do a network port as well (they are not expensive). I believe we would regret it later if we didn't. I have 5 PCs in our house --- not a single one of them has a printer port. Granted, one of the mother boards is printer port capable, but it didn't even come with a connector for it. At work people like the networked machines because they can program them from any machine. The software can be tricked into working over a network (if it doesn't have the built in capability to do so). In fact, our 3D printer can be started from home. You certainly wouldn't want to do that with a CNC mill, but who knows how the machine could be modified in the future. So many things are networked these days --- it would be painful to come up with a design based solely on a printer port.
Often these work. I've not found them be as reliable. In hooking them (and some USB/serial) adapters up to oscilloscopes they seem to have a lot of jittery signals. I've found the Lantronix and IOAcces ones work well for serial.
Bump... I've been working on an Eagle 6.x version of the UHU schematic. Not that we are totally committed to the UHU servo drive system but with an up-to-date version, we hope to have bids from board houses on fabricating them... IF the UHU is what we center on.
For those of you proficient in Eagle, you know the most difficult tasks often involve finding the correct part in the *vast* libraries available. If the part does not exist in a library, you have to make it yourself. This is not at the same pain level as dental surgery but it does take some time to learn to do it well.
So FWIW, this project is moving at least on two fronts: Machine mechanical drawing development (in another thread), and a UHU update development (centered on this thread) for fabrication bidding purposes.
With a frame/gantry/table a builder could choose the drive system from one of dozens already out there. We are not trying to reinvent the wheel but just get the wheel as cheap as possible within the budget goals of the group.
What I have personally done? Put some minor skin in the game by purchasing 3 each of the EAS board version of the UHU. That is what the Excel spreadsheet at the top of this thread is based on. My personal parts stash is big enough I might get them running except for the UHU controller chip that must be obtained from the original designer.
Excellent, R. D.! Keep us posted if you get the boards working!
Glad we have someone with electronics knowledge on board. I could chip in for the controllers, PM me if interested.
Friend req sent Kevin. Need it for PM communications.